Added gitignore and fixed Dockerfile.

.gitignore

@@ -0,0 +1,44 @@
+># Environment files
+.env
+.env.*
+
+# Virtual environment
+myenv/
+venv/
+ENV/
+
+# Model checkpoints
+checkpoints/
+
+# Python
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+
+# IDE
+.idea/
+.vscode/
+*.swp
+*.swo
+.DS_Store
+
+# Logs
+*.log
+logs/

Dockerfile

@@ -1,43 +1,46 @@
-# Use NVIDIA CUDA base image
-FROM nvidia/cuda:12.1.0-runtime-ubuntu22.04 as base
+# Use Python 3.10 as base image for better compatibility with ML libraries
+FROM python:3.10-slim
 
-# Set working directory to /code (Hugging Face Spaces convention)
-WORKDIR /code
+# Set working directory
+WORKDIR /app
 
-# Install system dependencies
-RUN apt-get update && apt-get install -y \
-    python3.10 \
-    python3-pip \
-    git \
-    && rm -rf /var/lib/apt/lists/*
+# Install git and required system dependencies
+RUN apt-get update && \
+    apt-get install -y git && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
 
-# Install Python packages
+# Copy requirements first to leverage Docker cache
 COPY requirements.txt .
-RUN pip3 install --no-cache-dir -r requirements.txt
 
-# Install any additional dependencies needed for litgpt
-RUN pip3 install --no-cache-dir \
-    einops \
-    xformers \
-    bitsandbytes \
-    accelerate \
-    sentencepiece
+# Install Python dependencies
+RUN pip install --no-cache-dir -r requirements.txt
 
-# Copy the application code
+# Copy the rest of the application
 COPY . .
 
-# Create model directory structure
-RUN mkdir -p /code/checkout/meta \
-    /code/checkout/microsoft \
-    /code/checkout/mistralai
+# Create checkpoints directory
+RUN mkdir -p /app/checkpoints
+
+# The token will be passed during build time
+ARG HF_TOKEN
+ENV HF_TOKEN=${HF_TOKEN}
+
+# Download the Llama 2 model using litgpt
+# Only proceed if HF_TOKEN is provided
+RUN if [ -n "$HF_TOKEN" ]; then \
+        python -c "from huggingface_hub import login; from litgpt.cli import download; login('${HF_TOKEN}'); download('meta-llama/Llama-2-3b-chat-hf', '/app/checkpoints')"; \
+    else \
+        echo "No Hugging Face token provided. Model will need to be downloaded separately."; \
+    fi
 
 # Set environment variables
-ENV PYTHONPATH=/code
 ENV LLM_ENGINE_HOST=0.0.0.0
-ENV LLM_ENGINE_PORT=7860
+ENV LLM_ENGINE_PORT=8001
+ENV MODEL_PATH=/app/checkpoints/meta-llama/Llama-2-3b-chat-hf
 
 # Expose the port the app runs on
 EXPOSE 8001
 
 # Command to run the application
-CMD ["python3", "main.py"]
+CMD ["python", "main/main.py"]